Key points are not available for this paper at this time.
with handheld dumbbells at 20%, 25% and 30% of body mass. Variables were compared across five integration methods using repeated measures Bland-Altman and two-way repeated measures ANOVA analyses (α = 0.05), with combined forward and backward integration serving as the criterion. Backward integration and after adjusting at the dumbbells release agreed with the criterion, while forward integration and adjusting at the bottom position did not. The dumbbell release point identified using 3D motion capture (criterion) was also compared to estimates derived from force platform data (vGRF method). The vGRF method identified the dumbbell release point in delay of 3D motion capture, with limits of agreement (LOA) between -0.17 and 0.03 s across conditions. These methods should not be used interchangeably; rather, we recommend that the vGRF method be used in situations whereby only force platforms are available, and that it is combined with forward and backward integration techniques.
Bright et al. (Sun,) studied this question.